Update README.md

README.md

@@ -18,7 +18,7 @@ We have finetuned [mesolitica/malaysian-mistral-7b-32k-instructions-v4](https://
 
 ### Training Details
 
-Overall, solely training on the [Boolq-Malay](https://huggingface.co/datasets/wanadzhar913/boolq-malay) dataset (comprised of both Malay and English versions of the original [Boolq](https://huggingface.co/datasets/google/boolq) dataset), we use the following training parameters and obtain the following training results:
+Overall, solely training on the [Boolq-Malay](https://huggingface.co/datasets/wanadzhar913/boolq-malay) dataset (comprised of both Malay and English versions of the original [Boolq](https://huggingface.co/datasets/google/boolq) dataset) and Google Colab's A100 GPU (40GB VRAM), we use the following training parameters and obtain the following training results:
 
 - **No. of Epochs:** 0.504
 - **Per Device Train Batch Size:** 4
@@ -63,9 +63,9 @@ nf4_config = BitsAndBytesConfig(
     bnb_4bit_compute_dtype=getattr(torch, TORCH_DTYPE)
 )
 
-tokenizer = AutoTokenizer.from_pretrained('wanadzhar913/malaysian-mistral-llmasajudge-v3')
+tokenizer = AutoTokenizer.from_pretrained('wanadzhar913/malaysian-mistral-llmasajudge-v2')
 model = AutoModelForCausalLM.from_pretrained(
-    'wanadzhar913/malaysian-mistral-llmasajudge-v3',
+    'wanadzhar913/malaysian-mistral-llmasajudge-v2',
     use_flash_attention_2 = True,
     quantization_config = nf4_config
 )